Your funnel baseline

One page that ends the fog

This is the lesson where things stop being abstract.

By the end of it you'll have a single page saying: how many people encounter your business each month, how many of them do something, how many buy, and how many come back. Four numbers and the rates between them.

That page is the most useful thing you will make in this module, because every module after this one is an attack on the weakest number on it. Module 4 attacks the rate from visit to enquiry. Module 6 attacks the top. Module 8 attacks the bottom. Without the baseline you can't choose where to aim, and — just as important — you can't prove afterwards that anything you did worked.

You will not have precise numbers. Nobody does. Estimates, clearly labelled as estimates, are the correct answer today.

Funnel arithmetic, and why small rates matter so much

Each stage is the previous stage multiplied by a conversion rate.

1,900 visitors × 2.2% = 41 trials 41 trials × 22% = 9 customers

The thing worth internalising: rate improvements compound, and they're usually cheaper than volume improvements. Doubling the visitor→trial rate from 2.2% to 4.4% doubles customers without a single extra visitor. Doubling visitors requires new content, new rankings, or new money — months of work, or a budget.

That's not an argument that rates are always the better target. It's an argument that you should look at both, and most people only ever look at volume.

Leak or drought?

Two different problems that feel identical from the inside.

A drought is a volume problem: not enough people find you. The rates are fine; the top of the funnel is empty.

A leak is a conversion problem: people arrive and don't act. The top is fine; something between stages is losing them.

They need opposite medicine. Pour more traffic into a leaky funnel and you waste most of it. Optimise conversion on a funnel nobody enters and you've spent a month improving the yield on nothing.

Most small businesses assume drought, because "we need more customers" naturally becomes "we need more people to see us". Sometimes that's right. Often it isn't, and the arithmetic tells you which.

Rough norms — with a serious warning

To judge whether a rate is bad, you need something to compare it to. Here are rough working ranges:

The warning matters more than the table. These are not standards, they're orientation. They vary enormously by industry, price point, traffic source and country. A €5,000 service converts differently from a €50 one. Branded traffic converts several times better than cold traffic.

Use them to notice something far out of range — an e-commerce site converting at 0.2%, or a service business closing 3% of enquiries. Do not use them to conclude you're "average and therefore fine". Your own trend over time is a better benchmark than anyone else's average, which is why today's baseline matters.

Three businesses, three different diagnoses

The same method produces genuinely different answers. This is the part worth studying.

Storkflow — a drought, with permission to spend.

1,900 visitors → 41 trials (2.2%) → 9 customers (22%) Marketing spend €600/month → CAC = €600 ÷ 9 = €67 CLV from Lesson 1.5 ≈ €1,020

Lena assumed her problem was that nothing was working. The arithmetic says otherwise: 2.2% visitor→trial and 22% trial→paid are both inside normal ranges. Her rates are fine. Her top is small.

But look at the last two lines. Every customer costs €67 to acquire and returns about €1,020 of margin. That's a ratio of roughly 15:1, when the working convention is 3:1. Lena isn't failing at marketing. She is drastically underspending on it, and the economics give her permission to spend perhaps three times more.

That is a completely different conversation to have with her managing director than "our marketing isn't working" — and she could only have it because she did this arithmetic.

Herzog Physio — a leak hiding behind a drought.

~800 website visitors → 6 online enquiries (0.75%) → most become patients Meanwhile: ~45 new patients a month arrive, ~30 by GP referral, ~12 by word of mouth

Two things are true at once. 800 visitors is small — a drought. But 0.75% of visitors enquiring is also well below what a local service site should manage, and there's an obvious reason: the website has no online booking, only a contact form going to an inbox Tomas checks weekly.

There's a third, larger observation. Forty-five patients arrive monthly and only about six come through the website at all. His digital funnel is almost irrelevant to his current business. That's not a reason to ignore it — it's the reason his growth is capped at however many GPs remember him. But it does mean he should be honest that he's building a channel, not fixing one.

Ostara Ceramics — the rates are fine, the risk is elsewhere.

6,500 sessions → 91 orders (1.4%) → 19% ever order again

1.4% is at the low end of the e-commerce range but not alarming. The number that should worry Mateus is 19% returning customers, on a product people plausibly buy repeatedly — you buy one bowl, then a set, then a gift. And the whole 6,500 sessions depend on a single algorithm they don't control.

His funnel doesn't have one leak. It has a hole at the bottom and a single point of failure at the top.

Hypothesis discipline

Spotting the weak number is the easy half. The lesson's actual demand is this: every leak observation must end in a sentence of the form —

"I believe [this] is happening because [reason], and I could test it by [cheap test]."

Compare:

❌ "Our website needs improving." ✅ "I believe visitors don't enquire because there's no way to book without phoning during working hours. I could test it by adding an online booking link and watching enquiries for four weeks."

The second one can be wrong, which is what makes it useful. Modules 2 through 4 will either confirm it or kill it, and either outcome is progress.

Now build yours.

---